More like this

More like this functionality.

Usage example

Sample document

from django.conf import settings

from django_elasticsearch_dsl import Document, Index, fields
from django_elasticsearch_dsl_drf.compat import KeywordField, StringField
from django_elasticsearch_dsl_drf.analyzers import edge_ngram_completion

from books.models import Book

from .analyzers import html_strip

INDEX = Index(settings.ELASTICSEARCH_INDEX_NAMES[__name__])

# See Elasticsearch Indices API reference for available settings
INDEX.settings(
    number_of_shards=1,
    number_of_replicas=1,
    blocks={'read_only_allow_delete': False}
)


@INDEX.doc_type
class BookDocument(Document):

    # ID
    id = fields.IntegerField(attr='id')

    title = StringField(
        analyzer=html_strip,
        fields={
            'raw': KeywordField(),
            'suggest': fields.CompletionField(),
            'edge_ngram_completion': StringField(
                analyzer=edge_ngram_completion
            ),
            'mlt': StringField(analyzer='english'),
        }
    )

    description = StringField(
        analyzer=html_strip,
        fields={
            'raw': KeywordField(),
            'mlt': StringField(analyzer='english'),
        }
    )

    summary = StringField(
        analyzer=html_strip,
        fields={
            'raw': KeywordField(),
            'mlt': StringField(analyzer='english'),
        }
    )

    # ...

    class Meta(object):
        """Meta options."""

        model = Book  # The model associate with this Document

    def prepare_summary(self, instance):
        """Prepare summary."""
        return instance.summary[:32766]

Sample view

from django_elasticsearch_dsl_drf.filter_backends import (
    FilteringFilterBackend,
    IdsFilterBackend,
    OrderingFilterBackend,
    PostFilterFilteringFilterBackend,
    SearchFilterBackend,
)
from django_elasticsearch_dsl_drf.viewsets import (
    DocumentViewSet,
    MoreLikeThisMixin,
)

from .serializers import BookDocumentSerializer

class BookMoreLikeThisDocumentViewSet(DocumentViewSet,
                                      MoreLikeThisMixin):
    """Same as BookDocumentViewSet, with more-like-this and no facets."""

    # ...

    document = BookDocument
    lookup_field = 'id'
    serializer_class = BookDocumentSerializer

    # ...

    filter_backends = [
        # ...
        FilteringFilterBackend,
        PostFilterFilteringFilterBackend,
        IdsFilterBackend,
        OrderingFilterBackend,
        SearchFilterBackend,
        # ...
    ]

    # More-like-this options
    more_like_this_options = {
        'fields': (
            'title.mlt',
            'summary.mlt',
            'description.mlt',
        )
    }

Sample request

http://localhost:8000/search/books-more-like-this-no-options/1007587/more_like_this/

Generated query

{
  "query": {
    "more_like_this": {
      "fields": [
        "title.mlt",
        "summary.mlt",
        "description.mlt"
      ],
      "like": {
        "_index": "book",
        "_id": "1007587",
        "_type": "book_document"
      }
    }
  },
  "from": 0,
  "size": 14,
  "sort": [
    "_score"
  ]
}

Options

Pretty much all Elasticsearch more-like-this options available. You might be particularly interested in the following:

  • min_term_freq
  • max_query_terms
  • unlike
  • stop_words